Contributors can still contribute patch files via JIRA as we’ve done for many 
years, no reason to change that.

Using github pull requests is for the day-to-day core development team to move 
quickly while still reviewing what we do. Every useful output of that effort 
will land in our git repository (the fixes / features themselves) or the 
mailing list (the review comments).

I share your general concern about depending on non-ASF infrastructure, but in 
this case it is transient (github can go offline or out of business after we’ve 
merged pull requests to git-wip-us without affecting us) and replaceable with 
some effort. I’ve seen Gerrit a bit and I like it, fwiw.

World-weary comments like "And I am tired none pay attention to such detail" is 
neither constructive nor fair, it would be nice to conduct at least one 
discussion on dev@ without them.

B.

On 19 Feb 2014, at 15:58, Benoit Chesneau <[email protected]> wrote:

> On Wed, Feb 19, 2014 at 4:48 PM, Benjamin Young <[email protected]> wrote:
>> On 2/19/14, 10:06 AM, Benoit Chesneau wrote:
>>> 
>>> On Wed, Feb 19, 2014 at 4:01 PM, Andy Wenk <[email protected]> wrote:
>>>> 
>>>> On 19 February 2014 15:56, Benoit Chesneau <[email protected]> wrote:
>>>>> 
>>>>> On Wed, Feb 19, 2014 at 3:47 PM, Andy Wenk <[email protected]> wrote:
>>>>>> 
>>>>>> On 19 February 2014 15:25, Robert Samuel Newson <[email protected]>
>>>>>> wrote:
>>>>>> 
>>>>>>> Yes. It's misleading for folks that stumble on it.
>>>>>>> 
>>>>>> +1
>>>>>> 
>>>>>> 
>>>>>>> On 19 Feb 2014, at 14:22, Noah Slater <[email protected]> wrote:
>>>>>>> 
>>>>>>>> Should we decommission our Review Board instance?
>>>>> 
>>>>> 
>>>>> well nobody really tried it ...
>>>>> 
>>>>> There is apparently some possibilities to bind automatically the
>>>>> review to review board, but not sure if it's feasible on apache. Also
>>>>> not sure It's the right tool, I preferred gerrit because it is abel to
>>>>> handle automatically the PRs from github.
>>>>> 
>>>>> It would be good to have the PR directly on apache. So people don't
>>>>> have to register to a privately held service just to review a code.
>>>>> Anyway. At first maybe people could really try the tool before taking
>>>>> any decision. I will try, myself when I am back in 2 weeks -if it's
>>>>> still there -. Waiting for my flight right now.
>>>>> 
>>>>> - benoit
>>>> 
>>>> 
>>>> maybe the fact that nobody tried to us it is a sign, that it is (at least
>>>> for now) not the right tool for the job? The efforts to use github for
>>>> reviews has for me (at least for now) shown, that this could lead into
>>>> the
>>>> right direction.
>>>> 
>>>> Save travels :)
>>>> 
>>> My concern is that it force people to go on a privately held service
>>> (and encouraging people to use it). Having notifications on the ml is
>>> awesome but not enough imo.
>> 
>> 
>> We'll always be dependent on something. The Apache Foundation's been OK with
>> the use of Github (afaik), so I don't see a problem with continuing. We're
>> not at risk of loosing code. We get a simpler, more familiar process for new
>> devs--something we need more of! And one of the simplest code review and
>> sharing available with no more maintenance time required from us or the
>> Foundation.
> 
> Beeing able to use github is not the same, as beeing forced to use
> github. (Off topic: I choose to be dependent on, not the contrary).
> 
> I am personally tired to see all this history going in the hand of a
> privately held company.. And I am tired none pay attention to such
> detail.
> 
> Anyway,  I would prefer to let the choice to people and having a
> channel between the tools. If we also get the diff from the PRs in the
> mailing list and if people are able to answer to them without needing
> to open an account on github, then all my concerns are gone.
> 
> - benoit

Reply via email to